In various cloud-computing systems, such as carrier-grade platform-as-a-service (cPaaS) systems, customers are able to deploy an application on a distributed cloud. In some of these deployments various virtual machines (VMs) for managing and providing the application may be provisioned on hardware located in geographically disparate data centers. In some of these applications, these VMs are required to communicate with each other. Such communication is typically via a public network such as the Internet. As such, in many systems, each VM may be given a public IP address to use in communicating with other application VMs, application managers, cloud management systems, customers, and other devices.